PyNN API

Allow the use of non standard synapses at the NEST level, and allow of non properly supported models. This was not really possible before to use all neuron types, and same with nativesynapses

fixed: Moved to  https://github.com/NeuralEnsemble/PyNN/issues/248

The fix in r1311 was incorrect. The merging is across populations, not across MPI nodes.

fixed: moved to  https://github.com/NeuralEnsemble/PyNN/issues/246

fixed

Fix the get_data() function for assemblies when called in an mpi environment

the function write_data is not working in parallel, when called on a Assembly object. Thus brunel.py is crashing in parallel. I'll try to find a fix

In case of plastic synapses, or if attributes that may not only be weight and delay are saved, it could be interesting to have a particular pyNN format for saving those connections. Then, a FromConnection? connector, for example, could open this ... [More] connection file and recreate the network. This is not a lot of work. A simple n_attribute*n_pre*n_post matrix can do the job, as it can already be obtained with save(format='array'). The catch is that for reloading we also need to store the metadata of the synapse attributes that are saved. [Less]

Documentation updates: switch from Subversion to Git, updated various URLs.

added .settings to gitignore